Related Experiment Videos
Algorithms to determine event timing during normal walking using kinematic data
1Department of Kinesiology and Health Science, California State University, Sacramento, 6000 J Street, Sacramento, CA 95819-6073, USA. ahreljac@hhs4.hhs.csus.edu
Journal of Biomechanics
|May 16, 2000
Summary
New algorithms accurately predict heelstrike and toeoff times during walking using only kinematic data. This method offers improved resolution over visual inspection for gait analysis.
Area of Science:
- Biomechanics
- Gait Analysis
- Motion Capture
Background:
- Accurate identification of gait events like heelstrike and toeoff is crucial for clinical and research applications.
- Traditional methods often rely on force plates or visual inspection, which can be limiting in terms of accessibility and resolution.
Purpose of the Study:
- To develop and validate algorithms for predicting heelstrike and toeoff times using solely kinematic data.
- To assess the accuracy and reliability of these algorithms compared to force platform data.
Main Methods:
- Kinematic data from two subjects walking at various speeds were collected using a 60Hz system.
- Algorithms were developed to predict key gait event timings from the kinematic data.
- The predicted timings were compared against synchronized force platform recordings.
Main Results:
- The average absolute error for predicting heelstrike was 4.7ms.
- The average absolute error for predicting toeoff was 5.6ms.
- True average errors were +1.2ms for both events, indicating minimal systematic bias.
Conclusions:
- The proposed algorithms provide an easy, reliable method for determining walking event times from kinematic data.
- This approach significantly improves temporal resolution compared to visual inspection of video.
- The algorithms are compatible with various 2-D and 3-D kinematic data collection systems.